Sedentary lifestyles are a significant public health concern, accounting for over 200,000 deaths each year in the U.S. Over 60 percent of adults do not meet minimal health recommendations for physical activity, so it is imperative to better understand the factors that explain physical activity. Psychosocial variables account for no more than 30-40 percent of the variance in physical activity. Macro-environmental variables may have large influences on daily physical activity of most Americans, but environmental variables are rarely included in physical activity studies. Research in the transportation field shows that land use (urban design) variables are related to walking for transportation, but the impact on total physical activity is not known. Community environments appear to stimulate walking when there is high residential density, more mixed use of land (residential and commercial), greater connectedness of streets, and a better infrastructure for walking (sidewalks). U.S. cities are among the lowest in the world on these "walkability" variables as well as walking for transportation. We hypothesize that people living in the most "walkable" communities that are high on an index of these variables do more total and moderate intensity physical activity than those who live in communities low on the index. The proposed study is the first to examine objectively measured macro-environmental variables and their relation to objectively measured levels of physical activity. The study will be conducted in two regions of the U.S. (Portland, OR and Boston, MA). A sample of 2,400 adults (75 subjects X 32 census tracts), aged 20 to 65 years and balanced by gender, will be randomly selected from 16 "high-walkable" and 16 "low-walkable" census tracts, matched on socioeconomic status. The land use variables will be measured objectively using existing Geographic Information System (GIS) databases. Physical activity will be assessed via accelerometers in two different seasons. Commonly studied psychosocial correlates of physical activity as well as perceived environment variables will be concurrently evaluated. The investigators have extensive experience studying physical activity in populations, using objective physical activity measures, using GIS methodology, and analyzing complex data sets that involve clustering. This innovative study will use methodologies from health and transportation research traditions to broaden understanding of potential influences on total and moderate intensity physical activity to include pervasive macro- environmental variables. Results will inform public health, land use, and transportation policies that could have population-wide impacts on physical activity.